HAL Id: halshs-03364050https://halshs.archives-ouvertes.fr/halshs-03364050

Submitted on 4 Oct 2021

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Tax Reforms and Political FeasibilityFelix Bierbrauer, Pierre Boyer, Andrew Lonsdale, Andreas Peichl

To cite this version:Felix Bierbrauer, Pierre Boyer, Andrew Lonsdale, Andreas Peichl. Tax Reforms and Political Feasi-bility. 2021. �halshs-03364050�

IPP Policy Briefs

n◦74

September 2021

Felix Bierbrauer*Pierre C. Boyer*Andrew LonsdaleAndreas Peichl*

*Authors of the reference study.

www.ipp.eu

The Institut des politiques publiques (IPP) has beencreated by PSE and is developed through a scien-tific partnership between the Paris School of Eco-nomics (PSE) and theGroupe des écoles nationalesd’économie et de statistique (GENES). The aim ofthe IPP is to promote quantitative analysis andevaluation of public policy using cutting edge re-search methods in economics.

Tax Reforms and Political Feasibility

Questions linked to the design and implementation of redistributive tax policieshave occupied a growing position on the public agenda over recent years. More-over, the fiscal pressures brought upon by the current coronavirus crisis will ensurethat these issues maintain considerable political significance for years to come. Inlight of this importance, we present novel research on reforms of income tax sys-tems. Our approach shows that tax reforms wherein the changes in individual taxburdens are larger for taxpayers with higher incomes are of particular interest.We denote such reforms as “monotonic” and show that, under this condition, it ispossible to determine the “winners” and “losers” of a given tax reform. One canthen conclude whether the monotonic reform is politically feasible, depending onwhether a majority of individuals will benefit financially from the policy. An em-pirical analysis of tax reforms with a focus on the United States and France revealsthat past reforms have, by and large, beenmonotonic. Our approach therefore en-ables us to test whether a given tax system admits a politically feasible reform andhas direct policy relevance for the common types of taxation reforms undertakenby government authorities.

� Identifying the winners and losers of a tax reform and determining which subset ofvoters must “win” for a reform to be politically feasible (in the sense that a majorityof voters will benefit financially from the policy) is possible when reforms are “mono-tonic”.

� Monotonic reforms are such that individuals with higher incomes experience largerchanges in their tax burdens. Examples include: (i) a reform that involves tax cuts forall incomes, with larger cuts for larger incomes; and (ii) a reform that involves highertaxes for everyone, with increases that are larger for the rich.

� In practice, the vast majority of tax reforms are monotonic: among the 394 incometax reforms that took place from 2000-2016 in a panel of 33 OECD countries, 78%were monotonic reforms. In France in particular, 85% of income tax reforms since1915 have been monotonic.

� The expansion of earnings subsidies schemes (such as the Earned Income Tax Creditin the United States or the Prime d’activité in France) is politically feasible.

� Tax cuts that are larger for individuals with higher incomes can be politically feasibleif policy-makers work to include the median voter among their beneficiaries.

IPP Policy Brief n◦74Tax Reforms and Political Feasibility

Introduction

General Background. The design of redistributive taxpolicies is an evergreen in the public discourse. Shouldhigh-income households be taxed more heavily? Shouldlow-income households be supported by means of earn-ings subsidies (e.g., the Earned Income Tax Credit in theUS or the Prime d’activité in France)? What tax rate shouldthe middle-class face? These questions are of particularconcern since major transformations of tax systems haveoccurred in recent decades, including: (1) a significant de-cline in top income tax rates in many OECD countries(Piketty, Saez, and Stantcheva, 2011); (2) the introduc-tion and subsequent increase of earnings subsidies (suchas the two programs mentioned above for the US andFrance); and (3) the introduction of sharp progressivityin the middle of the income distribution, which has haddetrimental implications for work incentives.1

These developments are puzzling from a political econ-omy perspective. Consider the debate on inequality, aprominent topic inOECDcountrieswhere concerns aboutrising inequality (and its consequences) have become cen-tral in political discussions. In democratic countries, wemight expect rising inequality to be partially offset by anincrease in political support for redistribution (Acemogluet al., 2014). Indeed, this is precisely the prediction of theclassical analysis by Meltzer and Richard (1981): higherinequality – as measured by the gap between averageand median income – should lead to an increased politicalpressure to engage in redistributive taxation. That suchtheoretical predictions have not borne out in practice pro-vides just one illustration of the difficulties faced whenpredicting the outcomes of political competition over taxpolicies. These challenges stem from the fact that themultiple policy instruments which make up contemporarytax and transfer systems (such as tax rates, the tax base,loopholes, etc.) have dynamic and countervailing effects -as a result, we lack a comprehensive understanding of thepolitical and economic factors that influence the develop-ment of tax policies. Such complexities, however, provideample opportunity for research that addresses pressingquestions at the crossroads of politics and taxation.Politically Feasible Income Tax Reforms. In light of thetopic’s clear importance to today’s economic climate, wehave developed a novel framework to assess income taxreforms through the lens of political feasibility (Bierbrauer,Boyer, and Peichl, 2021). To do so, we define a politicallyfeasible tax reform as an adjustment to the tax sched-ule that leaves a majority of the population better off fi-nancially. This policy note provides an overview of the

1Evidence of this pattern for the US is presented in Bierbrauer, Boyer,and Peichl (2021). Similar observations apply to Germany (where theproblem is referred to as the “Mittelstandsbauch” or “middle-class belly”- see Böcking and Hove, 2016), the Netherlands (Jacobs, Jongen, andZoutman, 2017), or France (Bierbrauer and Boyer, 2018).

promising applicability that our approach holds for thedevelopment of income tax policies in OECD countries,which we illustrate through a handful of historical exam-ples shown below. Our research examines the politicalsupport for tax reforms in contrast to tax systems them-selves, due to the clear relevance that the status quo taxpolicy2 has for today’s policy debate. Several examplesillustrate this point. First, the status quo policy is impor-tant in practice as political proposals often refer directlyto it. A brief look at the campaign manifestos of Don-ald J. Trump for the 2016 US presidential election andFrançois Hollande for the 2012 French presidential elec-tion helps make this clear by revealing explicit referencesto existing tax schedules: In particular, the Trump planstated that “[t]he current number of brackets” would “bereduced from seven to three” (Bump, 2019), while Hol-lande proposed to “make the wealthiest French residentscontribute to the national effort by introducing an addi-tional tax rate of 45% for incomes above 150,000 euros”(Liberation, 2012).3 Second, think tanks and policy an-alysts usually present the impact of reform proposals ondifferent taxpayers with respect to the status quo (see, forexample, Fabre et al., 2020, or the analyses by the Tax andPolicy Center reviewed in Bierbrauer, Boyer, and Peichl,2021). Our approach therefore presents government au-thorities with a novel framework to assess the desirabilityof tax reforms and holds clear relevance to the processthrough which policy-makers typically conceive and dis-cuss these changes.Our Research. We demonstrate that “monotonic” tax re-forms - reforms such that changes in individual tax bur-dens are larger for taxpayers with higher incomes – areof particular interest both from a theoretical as well as apractical perspective.4 As we show below, the vast ma-jority of income tax reforms observed in the past, andproposed during political campaigns and legislative de-bates, have been monotonic. This is evident both from anobservational analysis of historical tax reforms in OECDcountries (with a particular attention given to France), aswell as an in-depth microsimulation analysis of major pastchanges to the US tax system.From a theory perspective, if a reform is monotonic then amajority of the population will be in support of the policyif and only if the voter with median income is among itsbeneficiaries. This result makes it easy to study the polit-ical feasibility of any monotonic tax reform, as one onlyneeds to determine whether the median income voter

2The status quo tax policy simply refers to the existing tax scheduleat a given point time.

3The original quotation, which we have translated from French, was“Je ferai contribuer les plus fortunés des Francais à l’effort national encréant une tranche supplémentaire de 45% pour les revenus supérieursà 150 000 euros par part”.

4Examples of monotonic tax reforms in practice are: (1) reforms thatinvolve tax cuts for all incomes, with larger cuts for larger incomes; and(2) reforms that involve higher taxes for everyone, with increases thatare larger for the rich.

2

IPP Policy Brief n◦74Tax Reforms and Political Feasibility

stands to benefit in order to know if a majority of voterswill also gain from (and thus be in favor of) the policy.5 Itis also informative of the direction we should expect taxreforms to follow: the median voter will be in favor of re-forms that either raise tax rates above her income, or thatlower tax rates up to (and including) her income.To determinewhether a given reform is politically feasible,we identify upper and lower tax rate thresholds. Mono-tonic tax cuts for below-median incomes, and monotonictax increases for above-median incomes, will leave a ma-jority of the population better off if an only if the post-reform tax schedule falls between these two thresholds.In other words, these limits – which we denote “Paretobounds” – characterize the availability for government au-thorities to adjust income tax rates in a politically feasibleway.6 In our analysis below, we display three cases of up-per and lower Pareto bounds formajor past reforms of theUS federal personal income tax schedule. We show thatthese governments had more leeway to reduce taxes on“the poor” than they ultimately took advantage of, whilethe scope for feasible tax increases on “the rich” provedless clear-cut.

Tax Reforms in Theory and Practice

Tax Reforms in OECD Countries. First, we take a broadoverview look at the annual changes of statutory tax ratesin 33 OECD countries for the years 2000–2016. TheOECD provides annual data on the statutory tax systemsof its member countries.7 In particular, for singles with-out dependents, it documents tax brackets and tax ratesfor labor income. We use this information to constructa tax function, which calculates an individual’s income taxobligations for a given calendar year. A reform takes placewhen this function changes from one year to the next andis classified as monotonic when the change of the tax bur-den is larger for individuals with higher incomes.Table 1 shows that 78% of the reforms in the samplewere monotonic. The remainder include reforms that aremonotonic either above or below themedian, and reformswith non-monotonicities that seem economically negligi-ble. We provide more specific examples in the supple-mentary material of Bierbrauer, Boyer, and Peichl (2021).

5This result follows from the fact that monotonic tax reforms do not,in theory, affect the “rank” of individual incomes.

6Note that our reasoningmimics that found in commondiscussions of“Laffer Bounds”. Laffer Bounds define the range of tax rates over whichgovernments can adjust the income tax schedule without decreasinggovernment revenues (since, as it is often argued, inefficiently high taxrates will reduce the base of taxable income significantly by discourag-ing economic activity, while inefficiently low tax rates will simply yieldlow tax revenues). Although the idea of an inverse U-shaped relation-ship between tax rates and government revenues is often attributed toArthur Laffer (an economic advisor under the Reagan administration),the French engineer Jules Dupuit already formalized these insights in anacademic article from 1844 (Dupuit, 1844).

7See: https://stats.oecd.org/index.aspx?DataSetCode=TABLE_I1.

Table 1: Monotonic tax reforms in a panel of 33 OECDcountries (2000-2016).

Total number of possible reforms: 528Total number of reforms: 394

Number of monotonic reforms: 309 (78%)Number of non-monotonic reforms: 85 (22%)

Notes: Table 1 is based on theOECDdatabase (Table I.1. Centralgovernment personal income tax rates and thresholds) discussedin Bierbrauer, Boyer, and Peichl (2021) - see footnote 7. “Totalnumber of possible reforms” corresponds to the number of yearscovered in the data, while “Total number of reforms" is simply thenumber of reforms enacted.

Tax Reforms in France. We now show that the tendencyfor monotonic tax reforms within the OECD can be ob-served when looking at the history of French tax reforms,using the Institut des politiques publiques (IPP) database.8

Table 2 reports the monotonicity of all French incometax reforms since the first application of the tax in 1916.Monotonic reforms are the usual type of reform observedin France, making up 85% of those identified in our data.In the whole of French history, we observe just 11 non-monotonic reforms.

Table 2: Monotonic tax reforms in France (1915-2016).

Total number of possible reforms: 100Total number of reforms: 76

Number of monotonic reforms: 65 (85%)Number of non-monotonic reforms: 11 (15%)

Notes: Table 2 is based on the IPP database mentioned previ-ously - see footnote 8. “Total number of possible reforms” cor-responds to the number of years covered in the data, while “Totalnumber of reforms" is simply the number of reforms enacted.

Tax Reforms in the United States. We also take an in-depthmicrosimulation analysis of all major federal incometax reforms in the US since 1964 and find that they were,by and large, monotonic. Monotonic tax cuts – i.e., largertax cuts for richer taxpayers – were the most prevalentreform type. The first row of Figure 1 provides three ex-amples of such reforms under the Reagan, Bush Jr. andTrump administrations. The second row of Figure 1 alsodisplays the findings of three income tax reforms that ledto higher taxes on the top decile of American taxpayersunder the Bush Sr., Clinton and Obama administrations.Regardless of the type of reform displayed in Figure 1,we see that changes in individual tax burdens are typicallylarger for higher-income deciles, with the most significantchanges observed among the top decile of taxpayers. Astep-by-step breakdown of our microsimulation method-ology is presented in Box 1.

8See: https://www.ipp.eu/baremes-ipp/. Historical accounts ofFrench income tax reforms can be found in Piketty (2001), Delalande(2011), and André and Guillot (2014).

3

IPP Policy Brief n◦74Tax Reforms and Political Feasibility

Figure 1: Average reform-induced change in tax liability (y-axis) among income deciles of American taxpayers (x-axis)(a) Reagan (1986)

-150

0-1

000

-500

050

0

1 2 3 4 5 6 7 8 9 10

(b) Bush Jr. (2003)

-500

0-4

000

-300

0-2

000

-100

00

1 2 3 4 5 6 7 8 9 10

(c) Trump (2017)

-400

0-3

000

-200

0-1

000

0

1 2 3 4 5 6 7 8 9 10

(d) Bush Sr. (1990)

-100

010

020

030

0

1 2 3 4 5 6 7 8 9 10

(e) Clinton (1993)-5

000

500

1000

1 2 3 4 5 6 7 8 9 10

(f) Obama (2012)

-200

00

2000

4000

6000

1 2 3 4 5 6 7 8 9 10

Notes: Figure 1 shows the average value of the change in tax liability by decile for six reforms of the US federal personal income tax:the Reagan Tax Reform Act of 1986, the Bush Jobs and Growth Tax Relief Reconciliation Act of 2003, the Trump Tax Cuts and JobsAct of 2017, the Omnibus Budget Reconciliation Acts of 1990 and 1993 done by the Bush Sr. and Clinton administrations respectively,and the Obama American Taxpayer Relief Act of 2012. The red line represents a quadratic fit. Deciles are computed based on pre-taxincome without capital gains while the tax base includes capital gains. All computations are on the individual level. The vertical linesshow different locations for the median voter: the dashed line to the left imputes non-filers to the tax return data, while the dashedline to the right accounts for differential turnout by income. The solid line in the middle represents both the original median in the dataas well as the one accounting for both modifications simultaneously.Source: Figure 2 in Bierbrauer, Boyer, and Peichl (2021), see also Box 1 below.

Our analysis of the United States is based on the NBER’s microsimulation model TAXSIM and tax return microdata. Specifically,we use the public use files of tax return microdata from the Statistics of Income division of the IRS, which is provided by the NBER(Feenberg, 2012). These data include all information reported on individual tax returns from 1960 to 2012, which are availablebiannually for the years 1960–1966 and annually for the years 1966–2012. We use TAXSIM to calculate income and payroll taxes,as well as tax credits.

We construct a counterfactual measure of the change in a taxpayer’s tax burden that is only due to the reform, holding all individualcharacteristics, including the person’s income, constant. Let T0 be the tax system just prior to a given reform and T1 be the taxsystem immediately after it. We observe an individual i’s pre-tax income yi

0, and all further characteristics relevant to compute theindividual’s tax burden in the base period. We then use TAXSIM to calculate the person’s tax payment T0(y

i0). To account for the

fact that T1 becomes effective in the future, we compute an inflation-adjusted version of yi0 that we denote by yi

0. Our measure ofthe reform-induced change of the person’s tax burden is then T1(y

i0)− T0(y

i0). To see whether the given reform was a monotonic

tax reform, we then rank individuals according to pre-tax income and investigate to what extent tax units with higher incomesexperience larger changes of their tax burden than individuals with lower incomes.

For robustness, we invoke alternative ways of determining themedian in the income distribution. First, there is themedian positionin the tax return data we are using. Second, we make a correction for non-filers, i.e., low-income households who do not submit atax declaration. The new median income is then poorer than the one in the original data. Third, for a political economy analysis,the median income among voters is relevant. Since richer individuals are more likely to turn out, the median voter is richer thanthe median taxpayer in our data. Taking into account both non-filers and differential turnout brings us coincidentally back to themedian position in our data, i.e., these effects neutralize each other.

Box 1 : Microsimulation MethodologyBox 1 : Microsimulation Methodology

4

IPP Policy Brief n◦74Tax Reforms and Political Feasibility

Reform Proposals and the 2016 US Elections. Does thefinding that tax reforms are, by and large, monotonic, ex-tend to tax reform proposals which are publicly debatedbut not enacted? Providing an answer faces the challengethat such reform proposals often remain vague, so thatresearchers have to make assumptions about the missingdetails. To avoid own judgment calls, we invoke the sys-tematic analysis of reform proposals in the United Statesthat is provided by the Tax Policy Center.9 Their analysiscovers 69 reform proposals for the federal personal in-come tax that were made in the period 2003–2019: someproposals were made during presidential campaigns andprimaries, while others were put forth by administrationsduring the legislative process but did not come into ef-fect. Our findings for all 69 proposals reveal that the largemajority of tax reform proposals are monotonic. Figure2 provides an illustration of the results for key proposalsmade during the 2016 US presidential campaign and pri-maries, showing the average change in tax liability (on they-axis) by income quintile (on the x-axis) for the policiespresented by four prominent electoral candidates. Theproposals by the two Democratic candidates were of the“tax increase on the rich” type while the Republican pro-posals were of the “tax cuts for everybody” type. Regard-less of the type of reform shown, each proposal in thisfigure involves larger changes in individual tax burdens forhigher-income quintiles.

Figure 2: Changes in tax liability (y-axis) by incomequintile (x-axis), 2016 US elections

-40,

000

-20,

000

020

,000

40,0

00

All FirstQuintile

SecondQuintile

MiddleQuintile

FourthQuintile

TopQuintile

Trump Revised Clinton Revised Cruz Sanders

Notes: Figure 2 shows the average value of the change in taxliability for reform proposals made during the 2016 presiden-tial campaign and primaries for the US federal personal incometax, by income quintile. The first column shows the overall taxchange.Source: Figure 4 in Bierbrauer, Boyer, and Peichl (2021), basedon Tax Policy Center data.

9The Tax Policy Center is a prominent non-partisan think tank locatedinWashington, D.C., known for its influential analyses of issues linked toAmerican tax policy. See https://www.taxpolicycenter.org/.

Political Feasibility of Reform Direction

Tax Reforms in theUnited States. Wenow turn our atten-tion to marginal income tax rates – i.e., the rate applied tothe last unit of income that an individual earns10 – as thedecision to earn additional income depends on how thisincome is taxed and/or subsidized. Looking at changes inmarginal tax rates is therefore necessary to assess the im-plications of a given tax reform for tax revenues. We studywhether US tax reforms in the past led to lower marginaltax rates below the median income, possibly in connec-tion with higher rates above the median, and, in any case,more pronounced progression over a range of middle in-comes. We argue that the introduction and subsequentexpansions of the Earned Income Tax Credit indeed led tolower marginal tax rates for low incomes and more pro-nounced progression for slightly higher incomes, with noshift toward higher tax rates for above-median incomes.To provide a more detailed explanation of these obser-vations, we derive upper and lower Pareto bounds. Asnoted above, these thresholds determine the range overwhich reforms toward higher marginal tax rates above themedian income, and toward lower marginal tax rates upto (and including) the median income, are politically fea-sible. The amount of “space” that governments have toundertake such adjustments depends on the behavioralresponse to taxation – i.e., the extent to which individualsreduce their economic activity following an increase in in-come tax rates. Whether a change of the tax system gen-erates large or small behavioral responses from taxpayersis captured by the elasticity of taxable income (ETI), andthis parameter has implications for the efficiency of thereform. For instance, if this elasticity is low and taxpay-ers face an increase in their tax rates, we can expect therevenue generated by the reform to be significant.Figure 3 shows the pre- and post-reform values (in blueand red, respectively) of the ratio T ′(y)/(1−T ′(y)), whichis a function of the effective marginal tax rate T ′(y).11 Italso displays the upper Pareto bound (in colored dashedlines) for different values of the ETI.The purpose of this analysis is to assess whether, andunder which conditions, past reductions in the US fed-eral income tax can be justified on the grounds of po-litical feasibility. This would be the case if the effectivemarginal tax ratio had initially crossed the upper bound ina given sub-figure (indicating an inefficiently high statusquo tax schedule), and was subsequently brought below

10In the vast majority of modern tax systems, tax rates are organizedby income “bracket” such that individuals pay different levels of taxesover different ranges of their incomes.

11Wepresent the effectivemarginal tax rates in this ratio format as theexpression allows for a straightforward determination of the ETI. The ef-fective marginal tax rate accounts for the impact of changes in govern-ment benefits, in addition to the effect of income tax, on an individual’searnings as their income increases.

5

IPP Policy Brief n◦74Tax Reforms and Political Feasibility

Figure 3: Upper Pareto Bounds(a) Reagan (1986)

0

25

50

75

100

125

150

Effe

ctiv

e T'

/(1-T

') (in

%)

P25

P50

P75

P90

P95

P99

0 25 50 75 100 125

(b) Bush Jr. (2003)

0

25

50

75

100

125

150

Effe

ctiv

e T'

/(1-T

') (in

%)

P25

P50

P75

P90

P95

P99

0 25 50 75 100 125 150 175 200 225 250

(c) Trump (2017)

0

25

50

75

100

125

150

Effe

ctiv

e T'

/(1-T

') (in

%)

P25

P50

P75

P90

P95

P99

0 25 50 75 100 125 150 175 200 225 250

Notes: Figure 3 shows the ratio T ′/(1 − T ′) of the effective marginal tax rates before (solid blue line) and after (solid red line) forthree major reforms of the US federal personal income tax: the Reagan Tax Reform Act of 1986, the Bush Jobs and Growth Tax ReliefReconciliation Act of 2003, and the Trump Tax Cuts and Jobs Act of 2017. The dashed lines are the upper Pareto bounds for fivedifferent values of the elasticity of taxable income: 0.25 (khaki), 0.4 (lavender), 0.5 (cranberry), 0.75 (teal), 1 (orange) and 1.25 (green).Vertical dashed lines show different percentiles of the income distribution.Source: Figure 8 in Bierbrauer, Boyer, and Peichl (2021), see also Box 1.

Figure 4: Lower Pareto Bounds(a) Bush Sr. (1990)

-125

-100

-75

-50

-25

0

25

Effe

ctiv

e T'

/(1-T

') (in

%)

P10

P25

0 5 10 15

(b) Clinton (1993)

-125

-100

-75

-50

-25

0

25

Effe

ctiv

e T'

/(1-T

') (in

%)

P10

P25

0 5 10 15

(c) Obama (2012)

-125

-100

-75

-50

-25

0

25

Effe

ctiv

e T'

/(1-T

') (in

%)

P10

P25

0 5 10 15 20 25

Notes: Figure 4 shows the ratio T ′/(1 − T ′) of the effective marginal tax rates before (solid blue line) and after (solid red line) forthree major reforms of the US federal personal income tax involving earning subsides (Earned Income Tax Credit): the Omnibus BudgetReconciliation Acts of 1990 and 1993 done by the Bush Sr. and Clinton administrations respectively, and the Obama American TaxpayerRelief Act of 2012. The dashed lines are the lower Pareto bounds for four different values of the elasticity of taxable income: 5(cranberry), 4 (teal), 3 (orange) and 2 (green). Vertical dashed lines show different percentiles of the income distribution.Source: Figure 9 in Bierbrauer, Boyer, and Peichl (2021), see also Box 1.

this threshold by the reform in question. The first reformin Figure 3, undertaken by the Reagan administration, in-volved larger tax cuts for richer taxpayers and can be ra-tionalized as being politically feasible for an ETI above 0.5,but not for lower values. The 2003 tax cuts from the BushJr. administration and the 2017 Trump tax plan are onlypolitically feasible for an ETI above 0.75. In other words,whether these policies left the majority of the popula-tion better off depends on the extent of the behavioralresponse to taxation.Figure 4 shows the pre- and post-reform values (in blueand red, respectively) of the ratio T ′(y)/(1 − T ′(y)) forthree additional reforms of the US federal income tax,along with the lower Pareto bound (in colored dashedlines) for different values of the ETI. Similar to before,we are interested in determining whether these policiesmoved the effective marginal tax rate ratio away from aninefficiently low tax schedule. Each reform in Figure 4gives rise to the same conclusion: the lower bound camenowhere close to the pre- or post-reform systems. Hence,

lower tax rates for “the poor” were politically feasible. Theintroduction and subsequent expansion of the Earned In-come Tax Credit from the mid-1970s onward went in thisdirection, but did not take advantage of the full scope forpolitically feasible earnings subsidy expansions.Tax Reforms in France. Our in-depth analysis of the UScase raises the question of whether comparable insightscan be drawn for France at the dawn of the 2022 Frenchpresidential election. Additional research extending thisframework to past reforms of the French tax schedule in-deed points to similar conclusions (Bierbrauer and Boyer,2018). In particular, whether or not the upper Paretobound is crossed by the initial tax schedule depends onthe estimated value of the ETI. Over the last decade, theFrench tax schedule has been situated closer to its corre-sponding upper bound than that of the United States forany chosen ETI estimate,12 though this does not change

12In other words, the French tax schedule presented less room for po-litically feasible tax increases (or, conversely, provided stronger justifica-tion for tax cuts on above-median incomes).

6

IPP Policy Brief n◦74Tax Reforms and Political Feasibility

the nature of our underlying interpretation. Moreover, asin the US case, looking at the lower Pareto bound revealsa sizeable gap between this threshold and both the pre-and post-reform tax schedules for reforms previously un-dertaken in France. These findings illustrate the promis-ing applicability of our framework for politically feasibletax reform analysis beyond the US context.

Conclusion

The current coronavirus crisis will put an unprecedentedpressure on public finances. Raising revenues will be apriority once the virus recedes, and political feasibilityand fairness issues will be crucial to avoid fiscal revolts asin the French Yellow Vests protests (Boyer et al., 2019;Boyer et al., 2020). Tax systems have often been re-designed after major historical events and our ability totake these requirements into account moving forward willbe severely tested. The research outlined in this policynote has identified an unambiguous capacity for politi-cally feasible expansions of earnings subsidy programs inthe United States and France, with a less clear-cut scopefor changes to top marginal tax rates in these countries.Given the prevalence of monotonic reforms across theOECD, future research can help extend this frameworkto a range of cross-national contexts. While conductingthese analyses, recent literature contributed to our under-standing of what people know and learn about tax poli-cies, and how their support for different policies is deter-mined (Stantcheva, 2020). Our approach can help furthercontribute to this agenda and inform the development oftax policies both during this critical period and into theforeseeable future.

AuthorsFelix J. Bierbrauer is a Professor at the University ofCologne - Center for Macroeconomic Research (CMR).Pierre C. Boyer is a Professor at the École polytechnique- CREST and the director of the “Democracy and Institu-tions” Program at IPP.Andrew Lonsdale is a research assistant at CREST.Andreas Peichl is a Professor at the LMU-Munich and ifoInstitute.

Reference Study

Bierbrauer, Felix J., Pierre C. Boyer, and Andreas Peichl(2021). “Politically feasible reforms of non-linear taxsystems”. American Economic Review 111.1, pp. 153–91.

The authors gratefully acknowledge the support ofthe Max Planck Institute in Bonn, the Investissementsd’Avenir (ANR-11-IDEX-0003/Labex Ecodec/ANR-11-LABX-0047), Institut des Politiques Publiques, and theDeutsche Forschungsgemeinschaft (DFG, German Re-search Foundation) under Germany’s Excellence Strategy– EXC 2126/1– 390838866.

References

Acemoglu, Daron, Suresh Naidu, Pascual Restrepo, andJames Robinson (2014). "Can democracy help with in-equality?" VoxEU.org.

André, Mathias and Malka Guillot (2014). "1914 - 2014:One Hundred Years of Income Tax in France". IPP PolicyBriefs n◦12.

Bierbrauer, Felix J. and Pierre C. Boyer (2018). "Politicallyfeasible reforms of non-linear tax systems". CEPR Discus-sion Paper 13059.

Bierbrauer, Felix J., Pierre C. Boyer, and Andreas Peichl(2021). “Politically feasible reforms of non-linear taxsystems”. American Economic Review 111.1, pp. 153–91.

Boyer, Pierre C., Thomas Delemotte, Germain Gauthier,Vincent Rollet, and Benoit Schmutz (2019). "Le territoiredes gilets jaunes". IPP Policy Briefs n◦39.

– (2020). “Les déterminants de la mobilisation des Giletsjaunes”. Revue économique 71.1, pp. 109–138.

Bump, Philip (2019). "Three years later, evaluating the 10laws Trump said he’d pass in his first 100 days". TheWash-ington Post.

Böcking, David and Anna van Hove (2016). "Hier hängt derMittelstandsbauch". Der Spiegel.

Delalande, Nicolas (2011). Les batailles de l’impôt. Consen-tement et résistances de 1789 à nos jours. Le Seuil.

Dupuit, Jules (1844). “De la mesure de l’utilité des travauxpublics”. Revue économique (1995) 10.2, pp. 55–94.

Fabre, Brice, Arthur Guillouzouic, Chloé Lallemand, andClaire Leroy (2020). "Budget 2020 : quels effets pour lesménages ?" IPP Policy Briefs n◦49.

Feenberg, Daniel (2012). Tax Model Data at the NBER(1962–2012). National Bureau of Economic Research.

7

IPP Policy Brief n◦74Tax Reforms and Political Feasibility

Jacobs, Bas, Egbert L.W. Jongen, and Floris T. Zoutman(2017). “Revealed social preferences of Dutch politicalparties”. Journal of Public Economics 156, pp. 81–100.

Liberation (2012). "Les 60 engagements de Hollande".

Meltzer, Allan H. and Scott F. Richard (1981). “A RationalTheory of the Size of Government”. Journal of PoliticalEconomy 89.5, pp. 914–927.

Piketty, Thomas (2001). Les Hauts Revenus en France auXXe siècle. Inégalités et redistributions, 1901-1998. Gras-set.

Piketty, Thomas, Emmanuel Saez, and Stefanie Stantcheva(2011). “Taxing the 1%:Why the top tax rate could be over80%”. VoxEU.org.

Stantcheva, Stefanie (2020). "Understanding Tax Policy:How do people Reason?" CEPR Discussion Paper 15216.

Notes IPPEditorial board: P. Boyer, A. Bozio, J. Grenet

Managing editors: P. Dutronc-Postel, A. Rain